I finally have a little bit of time to come back and program ... here is the version that I came up with.  I left the variables intact and separate from one another, but if you examine the results you can easily see how they can be combined. data have;    input edge to from flow exposure; cards; 1 1 2 10 10 2 2 1  5 10 3 3 4  5  5 4 4 3  5  5 5 5 6 10 10 6 6 5  5  5 7 7 8  5  5 8 1 5 10 10 ; data matches;    set have;    rename edge=matching_edge             to=from           from=to           flow=matching_flow           exposure=matching_exposure; run; data combined;    if _n_=1 then do;       declare hash h (dataset:  'matches');       h.definekey('from', 'to');       h.definedata('matching_edge', 'matching_flow', 'matching_exposure');       h.definedone();       call missing (matching_edge, matching_flow, matching_exposure);    end;    set have;    if h.check()=0 then h.find(); run; proc print; run; ... View more

You could use a macro variable to store the list of variables. %let varlist=col1-col3 var4 ; data long ;   set have;   array c &varlist ;   do over c ;     col=c;     output;   end;   drop &varlist ; run; ... View more

I guess you found a simpler solution and you should stick to that. But Macro Arrays are really just a designation, you can achieve that easily using a correct naming. * one dim macro array, same thing for n dim macro array, just separate the dims with _; %let X_0=2; * store here the max dim of the array; %let X_1=HEIGHT; * cell 1; %let X_2=WEIGHT; * cell 2; then traversing the array is just a matter of using the right %do cycle: %macro runit(MVAR_ARRAY); proc sort data=sashelp.class out=want;     by %do _IDX=1 %to &&&MVAR_ARRAY._0;          &&&MVAR_ARRAY._&_IDX          %end;          ;   run;   proc print data=want;     var %do _IDX=1 %to &&&MVAR_ARRAY._0;         &&&MVAR_ARRAY._&_IDX         %end;         ;   run; %mend runit; %runit(X); Cheers from Portugal. ... View more